Quantification of preferences in markets

In this paper we quantify agent preferences in a market. In our framework every agent has a utility level associated with each transaction, and we assume that the probability of a feasible market transaction increases with an increase in total utility. It is surprising to observe that this simple behavioral principle induces a usually unique probability measure that can be constructed by a fast numerical algorithm. This unusual combination of a rigorous model and a fast numerical algorithm makes it possible to construct a well-defined set of preferences that implies a set of observed commodity prices.